threfunds

Inside the strategies shaping global capital.

Investment trends for 2026 shift toward AI infrastructure and energy

The investment thesis for the back half of the decade has crystallized faster than most client mandates have caught up.

Investment trends for 2026 shift toward AI infrastructure and energy

The capital reordering is already in motion

The figures out of Morgan Stanley Research are large enough to be disorienting. Roughly $2.9 trillion in global data center construction costs is expected through 2028, and more than 80% of that spending is still ahead. Goldman Sachs projects US data center electricity demand climbing from 31 GW in 2025 to 41 GW in 2026, then to 66 GW in 2027, while Gartner forecasts a 27% jump in worldwide data center power consumption this year — reaching 132 GW, up from 104 GW. This is no longer a technology narrative. It is an industrial one, reshaping GDP growth, earnings, and the competitive positioning of utilities across the cycle.

For us in advisory work, the channel mix is what matters. Morgan Stanley identifies $1.4 trillion in hyperscaler cash flows, $200 billion in corporate debt issuance, $150 billion in securitized credit, and roughly $800 billion in private credit opportunity through asset-based finance and joint venture debt funding the 2025–2028 capex cycle. An additional $350 billion is expected from private equity, venture capital, and sovereign investors. That is a meaningful share of new private credit origination, and it sits squarely inside the conversations clients are now asking us to lead.

Electrons, not chips, are the bottleneck

We keep hearing about chip constraints, but the real pinch point is power density. AI-optimized racks pull 30 to over 100 kilowatts against 5 to 15 for traditional infrastructure, and that load has overwhelmed local grid capacity in several regions. Companies are being forced to sign direct power purchase agreements with renewable generators, invest in on-site generation, or delay projects outright. The timeline mismatch is what makes the trade actionable: data centers can move from groundbreaking to operation in 9 to 12 months, while new power plants take 2 to 5 years.

The capital follows the gap. The IEA expects $2.2 trillion to flow into clean energy in 2026, nearly double the amount directed to fossil fuels, building on a record $2.3 trillion in 2025 (up 8.1% from 2024). For clients carrying sustainability mandates, this is where the trade-offs become concrete. Capital is moving toward renewables not as a climate preference but as an industrial necessity, and the return profile reflects that. The framing matters: investors who treat it as an ESG story will underweight what is, structurally, a power-and-grid story with its own risk surface.

What surfaces in client conversations this quarter

A few things worth raising when you next sit down with a client. First, productivity is no longer a slide in a pitch deck. Morgan Stanley notes that 21% of S&P 500 companies now cite AI benefits, and AI adopters are seeing cash-flow margin expansion at roughly twice the global average. When clients ask whether AI exposure is "priced in," the honest answer is that measurable monetization is starting to show up in earnings — but the capex cycle still has years to run, which is precisely where private credit and infrastructure allocations earn their keep in a generational portfolio.

Second, keep geography honest. PE-VC activity in India reportedly fell 3% to $6.45 billion in Q2CY26 as sector demand shifted, according to one industry tracker, and U.S. Bank's Q3 2026 outlook flags solid growth persisting even as policy and geopolitics redirect flows. The AI-energy thesis is global but not uniform, and concentration risk looks meaningfully different in mid-2026 than it did two years ago. Advisors who treat the trend as a single trade, rather than a set of regional exposures with their own transition-risk profile, will find themselves on the wrong side of a client question at some point in the next 12 months.

The takeaway for our practices is straightforward. The AI infrastructure buildout is a generational capital story, and it is also a generational client conversation story. The advisors who frame it as a constraint-driven industrial cycle — rather than a technology trade — will find their clients better prepared for the volatility still ahead.